None
EN
Labeling tools are great, but what about quality checks?
['Labeling Tools Are Great', 'But What About Quality Checks', 'Jakub Piotr Cłapa']
Collabora Newsroom RSS Feed
Fixing around 3% of label errors improves the model performance by 2%, although exact results will depend on the dataset and task. Thanks to MLfix, even a big dataset like the Mapillary Traffic Sign Dataset could be fully verified and fixed by a single person over a few days of work. Labeling is a difficult cognitive task and accurate labels require a serious Quality Assurance (QA) process.